Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

predicting future wars and explaining past wars

1 view
Skip to first unread message

Xcott Craver

unread,
Jan 26, 2002, 10:15:50 PM1/26/02
to
James "Kibo" Parry <ki...@world.std.com> wrote:
>Archimedes Plutonium (plut...@willinet.net) wrote:
>>
>> Okay, let us build the TABLE of Wars of past history:
>>
>> 1940s WW2
>>
>> 1910s WW1
>>
>> 1870s Franco Prussian War, USA Civil War
>>
>> 1810s Napoleonic Wars
>
>Wow, what a complete list! Most of them are even in the right decade.

This is a curious example of fudging data. At first it looks
like 4 data points, 1 of which is fudged; but since he's talking
about the time span between wars, it's really 3 data points,
only 1 of which is _not_ fudged.

>> from which we see there is a pattern of 30 year intervals or
>> 60 year intervals for which a reproduction population can
>> regain an *overpopulation* and thus bid for ensuing war.

Only Archie Pu can provide such a complex theory to explain
a single data point. Plus two extra values he gets to choose
to match the theory!

>"Hundred Years War". I'm sorry, the name of the "Hundred Years War"
>has destroyed your whole theory.

Also, he forgot when the Irish invaded Canada in 1866.

I am absolutely not making that up. The invasion force started
in Buffalo, NY, and was the first act of the IRB, later to
evolve kinda into the IRA. America didn't intervene at first,
probably because they considered the whole thing intensely silly.

-S


0 new messages